ForewordIntroductionScopeNormative referencesTerms and definitionsSymbolsEstimates of the parameters in the basic modelRequirements for a precision experimentLayout of the experimentRecruitment of the laboratoriesPreparation of the materialsPersonnel involved in a precision experimentPanelStatistical functionsExecutive functionsSupervisorsOperatorsStatistical analysis of a precision experimentPreliminary considerationsTabulation of the results and notation usedCellsRedundant dataMissing dataOutliersOutlying laboratoriesErroneous dataBalanced uniform-level test resultsCollation of data and intermediate valuesOriginal test resultsCell means (Form B of Figure 2) Measures of cell spread (Form C of Figure 2) Corrected or rejected dataScrutiny of results for consistency and outliersApproaches for scrutiny of dataGraphical consistency techniqueNumerical outlier techniqueCochran’s testGrubbs’ testsOne outlying observationTwo outlying observationsApplication of the Grubbs’ testRepeated testing for outlying means or outlying data pointsAlternative outlier inspection and test methodsCalculation of the general mean and variancesMethod of analysisBasic dataNon-empty cellsCalculation of the general mean, ^ Calculation of variancesAlternative calculation methods for variancesDependence of the variances upon m Establishing a functional relationship between precision values, s, and the mean level, m Choice of functional relationshipFitting relationships I and IIFitting relationship IIIFitting relationship IVStatistical analysis as a step-by-step procedureReport to the panel and decisions to be taken by the panelReport by the statistical expertDecisions to be taken by the panelFull reportStatistical tablesNumber of laboratories required for an estimate of precision (informative)Alternative calculations of variance components (informative)Calculation from a one-way analysis of variance tableRestricted maximum likelihood (REML) calculationExamples of the statistical analysis of precision experiments (informative)Example 1: determination of the sulfur content of coal (several levels with no missing or outlying data)BackgroundMeasurement methodSourceDescriptionGraphical presentationOriginal dataComputation of cell means, ¯ Computation of standard deviations, Scrutiny for consistency and outliersComputation of ^, s and s Dependence of precision on m ConclusionsAlternative calculationExample 2: softening point of pitch (several levels with missing data)BackgroundMeasurement methodSourceMaterialDescriptionGraphical presentationsOriginal dataCell meansAbsolute differences within cellsScrutiny for consistency and outliersComputation of ^, s and s Dependence of precision on m ConclusionsAlternative calculationExample 3: thermometric titration of creosote oil (several levels with outlying data)BackgroundSourceMaterialDescriptionOriginal dataCell meansAbsolute differences within cellsScrutiny for consistency and outliersComputation of ^, s and s Dependence of precision on m Alternative calculationCalculation of critical values and indicators (informative)Calculation of critical values for the Cochran's testCalculation of critical values for Grubbs' testsOne outlying observationTwo outlying observationsCalculation of indicators for Mandel’s h and k statisticsMandel’s h statisticMandel’s k statisticBibliography